论文概要
研究领域: NLP
作者: Peiwen Li, Harshit Verma, Sophia Zhou
发布时间: 2025-11-13
arXiv: 2511.10475
中文摘要
编码智能体在持续演化的代码库上执行连续任务,需要在下一个任务到来之前带着编辑记录并写好记忆。为未知任务选择记忆,要在未来实用性与写入、读取成本之间权衡。我们提出 CoReMem(Coding-Repository Memory),一个成本内化框架,通过在下一任务上实测记忆价值来指导记忆设计与智能体适配。前瞻性片段估值比较同一继承仓库状态下、有无各轨迹片段的下一任务运行,构建围绕锚点(仓库位置与可执行检查)的紧凑记忆。在 ChainSWE 上,CoReMem 将 27B 智能体的解决率从 21.0% 提升至 25.7%,步数减少 17%,含记忆模块调用的总缓存感知 token 数减少 11%。
原文摘要
Coding agents tackle successive tasks on an evolving repository, carrying forward edits and writing memory before the next task is known. Selecting memory for unknown next tasks requires balancing future usefulness against the costs of writing and reading it. We introduce Coding-Repository Memory (CoReMem), a cost-internalizing framework that guides memory design and actor adaptation through measured memory value on the next task. Prospective segment valuation compares runs of the next task with and without individual trajectory segments from the same inherited repository state, guiding a compact memory built around anchors: repository locations and executable checks. On ChainSWE, CoReMem raises the 27B actor's resolve rate from 21.0% to 25.7%, with 17% fewer steps and 11% fewer total cach...
自动采集于 2026-11-15
#论文 #arXiv #NLP #小凯
讨论回复
加载中...正在加载回复...
智谱 GLM-5 已上线
我正在智谱大模型开放平台 BigModel.cn 上打造 AI 应用,智谱新一代旗舰模型 GLM-5 已上线,在推理、代码、智能体综合能力达到开源模型 SOTA 水平。